Why Organisations Keep Learning the Same Thing Twice

• The condition whose absence produces no event
• It is not that nothing is written down
• Three places it is lost inside the company
• Three places it is lost across the boundary
• What the loss produces
• The Nordic dimension
• The Structural Response
• Conclusion
• References

Three of the four conditions for innovation announce themselves when they are missing; the fourth stays quiet, and its absence is experienced as normal speed.

The condition whose absence produces no event

Bridgium’s research across two studies identifies four conditions that determine whether innovation moves: legitimacy, predictability, connectivity and innovation memory. Three of them are self-announcing. Where legitimacy is missing, proposals get rejected and someone can name the rejection. Where predictability is missing, people complain about not knowing what happens next. Where connectivity is missing, it is visible that two units are not talking to each other.

Innovation memory behaves differently. Where it is absent, nothing happens that anyone would report. Work proceeds, meetings are held, decisions are made. They are simply made again, by different people, at a cost the organisation has no line item for. The absence produces no event, so it generates no signal, so it never reaches an agenda. This article is about that fourth condition and about where the research finds it failing.

The evidence comes from two studies. How Innovation Happens is based on 28 semi-structured interviews with innovation leaders and senior practitioners in large Nordic and European enterprises, conducted between September and December 2025, in organisations ranging from roughly 5,000 to 150,000 employees across manufacturing, energy, telecommunications, banking, enterprise software, infrastructure, chemicals and construction. From Discovery to Practice extends the analysis across the organisational boundary, drawing on 48 interviews in the same period, 28 with corporate innovation leaders and 20 with founders and senior representatives of startups and SMEs.

The first study defines the condition plainly: innovation memory concerns whether experience from discussions, pilots and previous attempts is retained and reused. That definition contains the whole problem, because retention and reuse are different capabilities and most organisations have only attempted the first.

The academic literature has a name for the underlying phenomenon. A systematic review of empirical work on knowledge loss induced by member turnover, published in The Learning Organization, identifies staff departures as one of the most frequently reported causes of organisational knowledge loss, and distinguishes forgetting that an organisation chooses from forgetting that simply happens to it. The distinction is useful here. Deliberately discarding obsolete knowledge is management. Losing the reasoning behind a decision because the person who made it changed role is not a decision at all, and it is the more common case.

It is not that nothing is written down

The immediate objection from anyone running a large organisation is that everything is documented. Project files, pilot reports, decision logs, gate reviews, handover packs. This is usually true and it is beside the point.

What gets recorded is outcomes: what was decided, what was built, what the pilot produced, whether it passed. What rarely gets recorded is reasoning: why an option was rejected, which constraint turned out to be binding, what was wrong with the original framing of the problem, what the team believed at the start that turned out to be false. Outcomes are what reporting requires. Reasoning is what prevents repetition, and nobody is measured on it.

Kerr’s observation about rewarding one behaviour while hoping for another applies with unusual precision. A project manager is assessed on delivery. Writing down why a promising direction was abandoned takes an afternoon, benefits a team two years from now, appears in no target, and carries a small risk of looking like an admission. The rational choice is to close the project and move on, and that is what happens.

There is a second reason documentation does not solve it. Nonaka and Takeuchi’s distinction between tacit and explicit knowledge describes what actually accumulates during an implementation: which approvals move quickly, whose objection arrives late if they are not consulted early, which integration looked simple and was not. That knowledge is held in the practice of the people who did the work rather than in any artefact, and no document format captures it because nobody experiences it as information while it is being acquired. It leaves with the person, and the organisation does not notice because there was never a file to be missing.

Three places it is lost inside the company

The first study models innovation inside an organisation as three stages, drawing on Berger and Luckmann’s account of how shared reality is constructed: making ideas speakable, making them shared and actionable, and making innovation part of practice. Its central observation is that the main difficulties appear at the transitions between stages rather than within them. An idea may be articulated and remain isolated. A pilot may succeed technically and lose ownership during handover. Learning from an experiment may disappear before it can be reused.

Each of those transitions loses a different thing, and each has a different cost when the organisation next attempts something similar.

Table 1. Where innovation memory fails inside an organisation. Constructed from Bridgium, How Innovation Happens (2026), transitions between the three stages of innovation flow.

The third row is the expensive one, and the least visible. An experiment that produced a clear negative result is genuinely valuable: it removes an option and narrows the field. But a negative result is only valuable if it is retained, and a negative result is exactly the kind of finding nobody circulates.

Three places it is lost across the boundary

The second study applies the same four conditions to collaboration between corporations and external partners, and finds innovation memory failing at all three stages of that process as well.

At the entry stage, the report states that previous contacts, scouting attempts, challenge definitions and ecosystem learning are not retained or made searchable, so both corporates and startups repeatedly restart the matching process instead of building on what has already been learned. At the pilot stage, the requirement is to document learnings, decisions and feedback, and the interviews describe feedback loops that are weak and slow. At the integration stage, the corporate requirement is to maintain feedback loops, document implementation learning and share knowledge across teams, while the external partner is expected to capture operational feedback, iterate based on adoption realities and support knowledge transfer.

Table 2. Where innovation memory fails across the organisational boundary. Constructed from Bridgium, From Discovery to Practice (2026), sections 5 to 7 and Table 5.

Comparing the two tables produces the observation this article exists to make. Inside a company, a lost record is a cost to one party. Across a boundary it is a cost to two, and the two losses are not symmetrical.

The corporation has continuity of institution and no continuity of person: the sponsor changes role, and with them goes the understanding of why this particular external partner was worth the effort. The external partner has continuity of person and no institutional standing: the founder remembers everything and can act on none of it, because the person who needs to be persuaded has been replaced. So the party with the memory has no authority, and the party with the authority has no memory. The relationship restarts, and both sides experience the restart as the other side’s fault.

What the loss produces

Four mechanisms follow, each observable in the interviews.

• Repeated discovery. The same constraint is found again by a different team, at the same cost, and the finding is experienced as new. Nobody can identify this as a repetition, because identifying it would require the record that is missing.
• Negative results that do not travel. An option ruled out in one unit is proposed again in another. Where it is proposed by someone senior, it is often attempted again in full, because the argument against it exists only as one person’s recollection.
• Restarted relationships. At the boundary, both sides begin the qualification conversation from the beginning. Founders in the research describe roughly a year of exploration before a proof of concept is approved, and a sponsor change resets a meaningful part of it.
• Capacity that never accumulates. Cohen and Levinthal showed that the ability to use external knowledge depends on internal capacity built through prior related experience. An organisation that does not retain what it learned from its last three pilots does not become better at running pilots. It becomes an organisation that has run three pilots, which is not the same asset.

The Nordic dimension

Two regional features shape how this condition fails in Nordic organisations, and one of them is a direct consequence of a genuine strength.

Horizontal information flow works well here. Flat structures, low power distance and direct cross-unit contact mean that knowledge moves efficiently by conversation, and asking the person who did it last time is quick and socially unremarkable. One respondent describes the pattern precisely before naming its limit:

“Information moves quite well, but decisions are still made within units. That’s where things slow down.”
— Digital Innovation Leader · Telecommunications · Finland

An organisation where knowledge travels by conversation has little incentive to write anything down, because the informal route works and is faster. That is efficient right up to the moment the person leaves, at which point the organisation discovers that its memory was a network of individuals rather than a record. The knowledge-loss literature identifies exactly this: turnover is among the most reported causes of organisational knowledge loss, and the effect is sharpest where retention depended on people rather than on artefacts.

The second feature is market size. In a Nordic sector containing a limited number of large buyers and a limited number of specialised suppliers, the same organisations meet repeatedly. A corporation that has forgotten why a previous engagement with a particular company stalled will re-enter it with the same assumptions, and the supplier on the other side of the table remembers everything. In a large market, forgetting a counterparty is survivable because there is another one. In a small market, the counterparty is the same one, and the asymmetry is visible to them.

The Structural Response

Four responses, all cheap relative to the cost of repetition.

1. Record reasoning, not only outcomes. At each decision point, one paragraph: what was chosen, what was rejected, and what would have to change for the rejected option to become viable. This is the smallest artefact that prevents repetition, and it takes minutes rather than the hours a formal lessons-learned process consumes and rarely produces.
2. Make the negative result a deliverable. If an experiment rules something out, that is the output and it should be written and circulated as such. Where only successes are reported, the organisation is buying the expensive half of experimentation and discarding the cheap half.
3. Keep a searchable record of the front door. Who was spoken to, about which problem, what was concluded and why. The research is explicit that entry-stage memory is weak, and the consequence is an annual scouting effort that starts from zero each time.
4. Treat sponsor handover as a transfer of context. When the person carrying an external relationship changes role, what has to move is not the contact record but the reasoning: why this partner, what has been established, what remains open. Without it, the successor inherits a name and a calendar invitation.

Conclusion

Innovation memory is the least discussed of the four conditions and the only one that degrades silently. The others fail loudly enough to reach a management meeting. This one produces an organisation that is busy, competent, and slowly re-deriving things it already knew, while experiencing that as the normal pace of change.

Both studies point at the same remedy from different directions, and it is unglamorous. Write down the reasoning, not only the result. Circulate what was ruled out. Keep the front door searchable. Hand over context rather than contacts. None of this requires a system, a budget line or a transformation programme, which may be precisely why it does not get done: there is nothing to procure and nobody’s role expands as a result.

So the test is a narrow one. Take the last initiative your organisation stopped, and ask whether anyone outside the team that stopped it could find out today why it was stopped. If not, the organisation has not finished paying for it.

Both reports are open. How Innovation Happens examines the flow inside organisations:
bridgium-research.eu/innovation-report-2026
From Discovery to Practice extends the analysis across the organisational boundary:
bridgium-research.eu/startup-report-2026

References

1. Bridgium Research Team (2026). How Innovation Happens: Insights from Leading Enterprises in Times of Change. Illarionova N., Verlin K., Verlin A. Report
2. Bridgium Research Team (2026). From Discovery to Practice: How Corporate–Startup Collaboration Becomes Usable. Illarionova N., Verlin K., Verlin A. Report
3. Knowledge loss induced by organizational member turnover: a review of empirical literature, synthesis and future research directions (Part I). The Learning Organization, 30(2). Emerald
4. Knowledge loss induced by organizational member turnover: a review of empirical literature, synthesis and future research directions (Part II). The Learning Organization. ScienceDirect
5. Berger, P. L. and Luckmann, T. (1966). The Social Construction of Reality. New York: Doubleday. Publisher
6. Nonaka, I. and Takeuchi, H. (1995). The Knowledge-Creating Company. New York: Oxford University Press. Publisher
7. Cohen, W. M. and Levinthal, D. A. (1990). Absorptive Capacity: A New Perspective on Learning and Innovation. Administrative Science Quarterly, 35(1), 128–152. JSTOR
8. Kerr, S. (1975). On the Folly of Rewarding A, While Hoping for B. Academy of Management Journal, 18(4), 769–783. JSTOR
9. March, J. G. (1991). Exploration and Exploitation in Organizational Learning. Organization Science, 2(1), 71–87. JSTOR
10. Weick, K. E. (1995). Sensemaking in Organizations. Thousand Oaks, CA: Sage. Publisher
11. Burt, R. S. (1992). Structural Holes: The Social Structure of Competition. Cambridge, MA: Harvard University Press. Publisher
12. McKinsey & Company (2020). Breaching the great wall to scale. Read